|
|
|
|
Large population deployment Although population sizes in evaluations of biometric systems increase, they are still many magnitudes away from planned applications like ID cards including biometric features of the ID holder. This problem requires discussion of theoretical and practical aspects of biometric systems in applications of large scale such as infrastructures, receiver operating rates, authentication error rates. Other important aspects here are for example binning strategies for largely distributed biometric databases. In a practical network of biometric devices a client may require authentication from many biometric devices operating on the same modality (e.g. face, fingerprint, voice, etc.); with each of these devices being in differing geographical positions. These devices may stem from differing commercial vendors which employ completely different device dependent biometric templates. A topic of considerable interest to the biometric community is how to exchange tem-plates between different commercial biometric de-vices so that a client would only have to enroll “once” in the network. Some recent advances have been made in this area with the advent of BioAPI and the Common Biometric Exchange File Format (CBEFF). However, considerable work still needs to be performed to realize the exchange of device independent biometric templates. Scalability of integrated systems It has been shown that integrated authentication systems on-device matching may significantly increase the security levels in biometrics. By encapsulation of sensor, biometric template and the verification mechanism into an integrated device such as a chip card, BM systems can be applied for autonomous tasks such as private key management in public-key-infrastructures. However, highly miniaturized devices such as smart cards today have very limited computational resources while on the other hand, algorithms used in BM are rather demanding respect to computational resources. The workshop seeks to reveal theoretical and practical approaches in down-sizing of BM algorithms to resolve the conflict arising from the miniaturization and integration requirements. Synergetic combination of biometric devices It can be shown experimentally that where an ensemble of experts (i.e. BM devices) that are of suitable performance and diversity; the combination of these local scores/decisions in the appropriate manner will provide a better global decision than any of the local decisions in isolation. However, there exist only a few theoretical explanations for these experimental results. In this workshop we would like to gain a greater understanding of issues peculiar to the synergetic combination of practical BM devices. Additionally in the workshop we would like to promote discussion on the non-canonical benefits of combining BM devices. For example, local decisions made by individual BM devices are often made with little to no knowledge of the context of the current sample of a claimant. Incorporating the external knowledge privy to a distributed network of practical BM devices (e.g. time of day, emergency status, construction work, recency of last visit, etc.) one can make more informed decisions and combine decisions from different devices in a more effective manner A number of scientific contests which deal with the evaluation of biometrics in laboratory conditions have been initiated recently. With the expected deployment of BM devices in large populations, additional problems can arise; such as to which degree do biometric systems fulfill their specifications (e.g. failure to enrolments). It will be of interest to discuss principles of deducting conclusions from laboratory tests to real-world scenarios from results of such activities. Furthermore, reports from field evaluations will provide a more exhaustive view to benchmarking. Improvements to user acceptance While for governmental driven deployment of biometrics like in ID cards, the user acceptance of biometric methods is of no significant concern, in the area of convenience applications or use in small user groups, biometric systems will only work properly if users behave cooperatively. For many biometric methods quite severe animosities from potential users have to be noticed. The workshop seeks to look at recent scientific and practical work towards an increased acceptance of biometrics by users. Biometric systems that are put into production should be designed to operate over a longer period of time. Besides the requirement of durable system components, the effect of gradual changes in physical or behavioral traits of individuals over time. This leads to the problem of aging, which shall be discussed in the workshop. Where do current biometrics fail? This workshop would be open to work that attempts to elucidate upon where and what applications some biometrics fail under. Biometrics in Human-to-Computer Interfaces (HCI) In this workshop we are also interested in Biometrics in Human-to-Computer Interfaces (HCI). We would be very interested in show-casing new applications for biometrics that are outside their canonical applications (e.g. security applications). Such areas might include, but are not limited to, human-machine interface (e.g. for disabled people) or safety applications like driver/pilot monitoring.
Satellite Workshop to, T
|
